Starch is an important constituent of cereal grains and of flours, accounting for about 65-67% of the weight of the grain at maturity. It is produced in the amyloplast of the grain endosperm by the concerted action of a number of enzymes, including ADP-Glucose pyrophosphorylase (EC 2.7.7.27), starch synthases (EC 2.4.1.21), branching enzymes (EC 2.4.1.18) and debranching enzymes (EC 3.2.1.41 and EC 3.2.1.68) (Ball et al, 1996; Martin and Smith, 1995; Morell et al, 1995). Some of the proteins involved in the synthesis of starch can be recovered from the starch granule (Denyer et al, 1995; Rahman et al, 1995).
Most wheat cultivars normally produce starch containing 25% amylose and 75% amylopectin. Amylose is composed of large linear chains of α (1-4) linked α-D-glucopyranosyl residues, whereas amylopectin is a branching form of α-glycan linked by α (1-6) linkages. The ratio of amylose and amylopectin, the branch chain length and the number of branch chains of amylopectin are the major factors which determine the properties of wheat starch.
Starch with various properties has been widely used in industry, food science and medical science. High amylose wheat can be used for plastic substitutes and in paper manufacture to protect the environment; in health foods to reduce bowel cancer and heart disease; and in sports foods to improve the athletes, performance. High amylopectin wheat may be suitable for Japanese noodles, and is used as a thickener in the food industry.
Wheat contains three sets of chromosomes (A, B and D) in its very large genome of about 1010 base pairs (bp). The donor of the D genome to wheat is Triticum tauschii, and by using a suitable accession of this species the genes from the D genome can be studied separately (Lagudah et al, 1991).
There is comparatively little variation in starch structure found in wheat varieties, because the hexaploid nature of wheat prevents mutations from being readily identified. Dramatic alterations in starch structure are expected to require the combination of homozygous recessive alleles from each of the 3 wheat genomes, A, B and D. This requirement renders the probability of finding such mutants in natural or mutagenised populations of wheat very low. Variation in wheat starch is desirable in order to enable better tailoring of wheat starches for processing and end-user requirements.
Key commercial targets for the manipulation of starch biosynthesis are:                1. “Waxy” wheats in which amylose content is decreased to insignificant levels. This outcome is expected to be obtained by eliminating granule-bound starch synthase activity.        2. High amylose wheats, expected to be obtained by suppressing starch branching enzyme-II activity.        3. Wheats which continue to synthesise starch at elevated temperatures, expected to be obtained by identifying or introducing a gene encoding a heat-stable soluble starch synthase.        4. “Sugary types” of wheat which contain increased amylose content and free sugars, expected to be obtained by manipulating an isoamylase-type debranching enzyme.        
There are two general strategies which may be used to obtain wheats with altered starch structure:                (a) using genetic engineering strategies to suppress the activity of a specific gene, or to introduce a novel gene into a wheat line; and        (b) selecting among existing variation in wheat for missing (“null”) or altered alleles of a gene in each of the genomes of wheat, and combining these by plant breeding.However, in view of the complexity of the gene families, particularly starch branching enzyme I (SBE I), without the ability to target regions which are unique to genes expressed in endosperm, modification of wheat by combination of null alleles of several enzymes in general represents an almost impossible task.        
Branching enzymes are involved in the production of glucose α-1,6 branches. Of the two main constituents of starch, amylose is essentially linear, but amylopectin is highly branched; thus branching enzymes are thought to be directly involved in the synthesis of amylopectin but not amylose. There are two types of branching enzymes in plants starch branching enzyme I (SBE I) and starch branching enzyme II (SBE II), and both are about 85 kDa in size. At the nucleic acid level there is about 65% sequence identity between types I and II in the central portion of the molecules; the sequence identity between SBE I from different cereals is about 85% overall (Burton et al, 1995; Morell et al, 1995).
In cereals, SBE I genes have so far been reported only for rice (Kawasaki et al, 1991; Rahman et al, 1997). A cDNA sequence for wheat SBE I is available on the GenBank database (Accession No. Y12320; Repellin A., Nair R. B., Baga M., and Chibbar R. N.: Plant Gene Register PGR97-094, 1997). As far as we are aware, no promoter sequence for wheat SBE I has been reported.
We have characterised an SBE I gene, designated wSBE I-D2, from Triticum tauschii, the donor of the D genome to wheat (Rahman et al, 1997). This gene encoded a protein sequence which had a deletion of approximately 65 amino acids at the C-terminal end, and appeared not to contain some of the conserved amino acid motifs characteristic of this class of enzyme (Svensson, 1994). Although wSBE I-D2 was expressed as mRNA, no corresponding protein has yet been found in our analysis of SBE I isoforms from the endosperm, and thus it is possible that this gene is a transcribed pseudogene.
Genes for SBE II are less well characterised; no genomic sequences are available, although SBE II cDNAs from rice (Mizuno et al, 1993; Accession No. D16201) and maize (Fisher et al, 1993; Accession No. L08065) have been reported. In addition, a cDNA sequence for SBE II from wheat is available on the GenBank database (Nair et al, 1997; Accession No. Y11282); although the sequences are very similar to those reported herein, there are differences near the N-terminal of the protein, which specifies its intracellular location. No promoter sequences have been reported, as far as we are aware.
Wheat granule-bound starch synthase (GBSS) is responsible for amylose synthesis, while wheat branching enzymes together with soluble starch synthases are considered to be directly involved in amylopectin biosynthesis. A number of isoforms of soluble and granule-bound starch synthases have been identified in developing wheat endosperm (Denyer et al, 1995). There are three distinct isoforms of starch synthases, 60 kDa, 75-77 kDa and 100-105 kDa, which exist in the starch granules (Denyer et al, 1995; Rahman et al, 1995). The 60 kDa GBSS is the product of the wx gene. The 75-77 kDa protein is a wheat soluble starch synthase I (SSSI) which is present in both the soluble fraction and the starch granule-bound fraction of the endosperm. However, the 100-105 kDa proteins, which are another type of soluble starch synthase, are located only in starch granules (Denyer et al, 1995; Rahman et al, 1995). To our knowledge there has been no report of any complete wheat SSS I sequence, either at the protein or the nucleotide level.
Both cDNA and genomic DNA encoding a soluble starch synthase I of rice have been cloned and analysed (Baba et al, 1993; Tanaka et al, 1995). The cDNAs encoding potato soluble starch synthase SSSII and SSSIII and pea soluble starch synthase SSSII have also been reported (Edwards et al, 1995; Marshall et al, 1996; Dry et al, 1992). However, corresponding full length cDNA sequences for wheat have hitherto not been available, although a partial cDNA sequence (Accession No. U48227) has been released to the GenBank database.
Approach (b) referred to above has been demonstrated for the gene for granule-bound starch synthase. Null alleles on chromosomes 7A, 7D and 4A were identified by the analysis of GBSS protein bands by electrophoresis, and combined by plant breeding to produce a wheat line containing no GBSS, and no amylose (Nakamura et al, 1995). Subsequently, PCR-based DNA markers have been identified, which also identify null alleles for the GBSS loci on each of the three wheat genomes. Despite the availability of a considerable amount of information in the prior art, major problems remain. Firstly, the presence of three separate sets of chromosomes in wheat makes genetic analysis in this species extraordinarily complex. This is further complicated by the fact that a number of enzymes are involved in starch synthesis, and each of these enzymes is itself present in a number of forms, and in a number of locations within the plant cell. Little, if any, information has been available as to which specific form of each enzyme is expressed in endosperm. For wheat, a limited amount of nucleic acid sequence information is available, but this is only cDNA sequence; no genomic sequence, and consequently no information regarding promoters and other control sequences, is available. Without being able to demonstrate that the endosperm-specific gene within a family has been isolated, such sequence information is of limited practical usefulness.